Plant electrical signal-based method for identifying copper stress in lettuce

The healthy growth of crops is very important for itself, especially the heavy metal pollution in the growth environment is especially worth studying. The identification of heavy metal pollution in crops usually requires long-term morphological observation or physiological and biochemical experiments, which is time-consuming and labor-intensive. Addressing the limitations of traditional detection methods, which rely on a single signal feature and exhibit restricted classification capabilities, this study examined lettuce leaves subjected to copper ion stress alongside healthy lettuce leaves. The study systematically analyzed the evolutionary patterns of plant electrical signal characteristics across different stages of copper stress. Building upon this analysis, a novel lettuce copper stress identification method was proposed, integrating multi-wavelet entropy features with BP_Adaboost ensemble learning. Experiments revealed that lettuce exhibits a dynamic evolution of its electrophysiological response to copper stress, progressing through stages of “stress-transition-adaptation.” During the 2 h stress period, signal characteristics exhibited significant differences, with model accuracy peaking at 94.0%. Subsequently, during the 4 h transition period, accuracy declined to 87.3%. In the 6 h adaptation period, accuracy further decreased to 63.3% due to the restoration of physiological homeostasis. The integrated model outperformed both BP and SVM algorithms across all stages, demonstrating its effectiveness in capturing early stress features in plants and offering a novel approach for the early monitoring of heavy metal contamination in crops.